Qualitative Topological Coverage of Unknown Environments by Mobile Robots

This thesis considers the problem of complete coverage of unknown environments by a mobile robot. The goal of such navigation is for the robot to visit all reachable surfaces in an environment. The task of achieving complete coverage in unknown environments can be broken down into two smaller sub-tasks. The first is the construction of a spatial representation of the environment with information gathered by the robot's sensors. The second is the use of the constructed model to plan complete coverage paths. A topological map is used for planning coverage paths in this thesis. The landmarks in the map are large scale features that occur naturally in the environment. Due to the qualitative nature of topological maps, it is rather difficult to store information about what area the robot has covered. This difficulty in storing coverage information is overcome by embedding a cell decomposition, called slice decomposition, within the map. This is achieved using landmarks in the topological map as cell boundaries in slice decomposition. Slice decomposition is a new cell decomposition method which uses the landmarks in the topological map as its cell boundaries. It decomposes a given environment into non-overlapping cells, where each cell can be covered by a robot following a zigzag pattern. A new coverage path planning algorithm, called topological coverage algorithm, is developed to generate paths from the incomplete topological map/slice decomposition, thus allowing simultaneous exploration and coverage of the environment. The need for different cell decompositions for coverage navigation was first recognised by Choset. Trapezoidal decomposition, commonly used in point-to-point path planning, creates cells that are unnecessarily small and inefficient for coverage. This is because trapezoidal decomposition aims to create only convex cells. Thus, Choset proposed boustrophedon decomposition. It introduced ideas on how to create larger cells that can be covered by a zigzag, which may not necessarily be convex. However, this work is conceptual and lacking in implementation details, especially for online creation in unknown environments. It was later followed by Morse decomposition, which addressed issues on implementation such as planning with partial representation and cell boundary detection with range sensors. The work in this thesis was developed concurrently with Morse decomposition. Similar to Morse decomposition, slice decomposition also uses the concepts introduced by boustrophedon decomposition. The main difference between Morse decomposition and slice decomposition is in the choice of cell boundaries. Morse decomposition uses surface gradients. As obstacles parallel to the sweep line are non-differentiable, rectilinear environments cannot be handled by Morse decomposition. Also, wall following on all side boundaries of a cell is needed to discover connected adjacent cells. Therefore, a rectangular coverage pattern is used instead of a zigzag. In comparison, slice decomposition uses topology changes and range sensor thresholding as cell boundaries. Due to the use of simpler landmarks, slice decomposition can handle a larger variety of environments, including ones with polygonal, elliptical and rectilinear obstacles. Also, cell boundaries can be detected from all sides of a robot, allowing a zigzag pattern to be used. As a result, the coverage path generated is shorter. This is because a zigzag does not have any retracing, unlike the rectangular pattern. The topological coverage algorithm was implemented and tested in both simulation and with a real robot. Simulation tests proved the correctness of the algorithm; while real robot tests demonstrated its feasibility under inexact conditions with noisy sensors and actuators. To evaluate experimental results quantitatively, two performance metrics were developed. While there are metrics that measure the performance of coverage experiments in simulation, there are no satisfactory ones for real robot tests. This thesis introduced techniques to measure effectiveness and efficiency of real robot coverage experiments using computer vision techniques. The two metrics were then applied to results from both simulated and real robot experiments. In simulation tests, 100% coverage was achieved for all experiments, with an average path length of 1.08. In real robot tests, the average coverage and path length attained were 91.2% and 1.22 respectively.

[1]  Thomas Ertl,et al.  Computer Graphics - Principles and Practice, 3rd Edition , 2014 .

[2]  J. Hopcroft,et al.  Algorithm 447: efficient algorithms for graph manipulation , 1973, CACM.

[3]  Viii Supervisor Sonar-Based Real-World Mapping and Navigation , 2001 .

[4]  Ronald C. Arkin,et al.  An Behavior-based Robotics , 1998 .

[5]  Tom Duckett,et al.  Mobile robot self-localisation and measurement of performance in middle-scale environments , 1998, Robotics Auton. Syst..

[6]  Alexander Zelinsky,et al.  Planning Paths of Complete Coverage of an Unstructured Environment by a Mobile Robot , 2007 .

[7]  Alexander Zelinsky,et al.  A mobile robot exploration algorithm , 1992, IEEE Trans. Robotics Autom..

[8]  Shigeru Sakamoto Mechanical planning and actual test results of a robot for painting the exterior walls of high-rise buildings , 1990, Adv. Robotics.

[9]  Andreas Kurz Constructing maps for mobile robot navigation based on ultrasonic range data , 1996, IEEE Trans. Syst. Man Cybern. Part B.

[10]  Philip D. Loewen Optimal Control Via Nonsmooth Analysis , 1993 .

[11]  Rolf Klein,et al.  On the Competitive Complexity of Navigation Tasks , 2000, Sensor Based Intelligent Robots.

[12]  Anthony Stentz,et al.  Multi-robot exploration controlled by a market economy , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[13]  Elon Rimon,et al.  Spanning-tree based coverage of continuous areas by a mobile robot , 2001, Proceedings 2001 ICRA. IEEE International Conference on Robotics and Automation (Cat. No.01CH37164).

[14]  Rodney A. Brooks,et al.  Elephants don't play chess , 1990, Robotics Auton. Syst..

[15]  R. A. Brooks,et al.  Intelligence without Representation , 1991, Artif. Intell..

[16]  S. Hyakin,et al.  Neural Networks: A Comprehensive Foundation , 1994 .

[17]  Uwe R. Zimmer,et al.  Robust world-modelling and navigation in a real world , 1996, Neurocomputing.

[18]  Dolores Canamero,et al.  Designing emotions for activity selection in autonomous agents , 2003 .

[19]  Maja J. Mataric,et al.  Sold!: auction methods for multirobot coordination , 2002, IEEE Trans. Robotics Autom..

[20]  Benjamin Kuipers,et al.  A Robust, Qualitative Method for Robot Spatial Learning , 1988, AAAI.

[21]  Howie Choset,et al.  Path Planning for Robotic Demining: Robust Sensor-Based Coverage of Unstructured Environments and Probabilistic Methods , 2003, Int. J. Robotics Res..

[22]  Nixon,et al.  Feature Extraction & Image Processing , 2008 .

[23]  Frédéric Lerasle,et al.  Topological navigation and qualitative localization for indoor environment using multi-sensory perception , 2002, Robotics Auton. Syst..

[24]  Amol Dattatraya Mali On the behavior-based architectures of autonomous agency , 2002, IEEE Trans. Syst. Man Cybern. Part C.

[25]  Elisha Sacks,et al.  Path planning for planar articulated robots using configuration spaces and compliant motion , 2003, IEEE Trans. Robotics Autom..

[26]  Sylvia C. Wong,et al.  A topological coverage algorithm for mobile robots , 2003, Proceedings 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003) (Cat. No.03CH37453).

[27]  J. O'Keefe,et al.  The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. , 1971, Brain research.

[28]  Kaspar Althoefer,et al.  Pipe inspection using a laser-based transducer and automated analysis techniques , 2003 .

[29]  Jin Bae Park,et al.  Navigation of cleaning robots using triangular-cell map for complete coverage , 2003, 2003 IEEE International Conference on Robotics and Automation (Cat. No.03CH37422).

[30]  Maja J. Matarić,et al.  A General, Local Algorithm for Robot Formations , 2001 .

[31]  Norbert Elkmann,et al.  Innovative service robot systems for facade cleaning of difficult-to-access areas , 2002, IEEE/RSJ International Conference on Intelligent Robots and Systems.

[32]  Vladimir J. Lumelsky,et al.  A terrain-covering algorithm for an AUV , 1996, Auton. Robots.

[33]  Hobart R. Everett,et al.  Sensors for Mobile Robots: Theory and Application , 1995 .

[34]  Cristina P. Santos,et al.  Control of a robot painting system using the multi-resolution architectural principle-a summary , 1997, ISIE '97 Proceeding of the IEEE International Symposium on Industrial Electronics.

[35]  Keiji Nagatani,et al.  Topological simultaneous localization and mapping (SLAM): toward exact localization without explicit localization , 2001, IEEE Trans. Robotics Autom..

[36]  M. V. Velzen,et al.  Self-organizing maps , 2007 .

[37]  Shuzhi Sam Ge,et al.  Dynamic Motion Planning for Mobile Robots Using Potential Field Method , 2002, Auton. Robots.

[38]  Bruce A. MacDonald,et al.  An evaluation of the sequential Monte Carlo technique for simultaneous localisation and map-building , 2003, 2003 IEEE International Conference on Robotics and Automation (Cat. No.03CH37422).

[39]  Elon Rimon,et al.  Spiral-STC: an on-line coverage algorithm of grid environments by a mobile robot , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[40]  Howie Choset,et al.  Complete sensor-based coverage with extended-range detectors: a hierarchical decomposition in terms of critical points and Voronoi diagrams , 2001, Proceedings 2001 IEEE/RSJ International Conference on Intelligent Robots and Systems. Expanding the Societal Role of Robotics in the the Next Millennium (Cat. No.01CH37180).

[41]  Francesco Mondada,et al.  Mobile Robot Miniaturisation: A Tool for Investigation in Control Algorithms , 1993, ISER.

[42]  Michael Mayo,et al.  Using Context to Solve the Correspondence Problem in Simultaneous Localisation and Mapping , 2004, PRICAI.

[43]  Jeremy G. Siek,et al.  The Boost Graph Library - User Guide and Reference Manual , 2001, C++ in-depth series.

[44]  Steven L. Tanimoto The Elements of Artificial Intelligence Using Common Lisp , 1995 .

[45]  Anita M. Flynn,et al.  Mobile robots: inspiration to implementation , 1993 .

[46]  Sylvia C. Wong,et al.  Complete Coverage by Mobile Robots Using Slice Decomposition Based on Natural Landmarks , 2004, PRICAI.

[47]  Sebastian Thrun,et al.  Robotic mapping: a survey , 2003 .

[48]  Dennis G. Zill,et al.  Advanced Engineering Mathematics , 2021, Technometrics.

[49]  H. Barlow Vision: A computational investigation into the human representation and processing of visual information: David Marr. San Francisco: W. H. Freeman, 1982. pp. xvi + 397 , 1983 .

[50]  Rodney A. Brooks,et al.  A Robust Layered Control Syste For A Mobile Robot , 2022 .

[51]  G. Swaminathan Robot Motion Planning , 2006 .

[52]  Ulrich Nehmzow,et al.  Animal and robot navigation , 1995, Robotics Auton. Syst..

[53]  Günther Schmidt,et al.  Building a global map of the environment of a mobile robot: the importance of correlations , 1997, Proceedings of International Conference on Robotics and Automation.

[54]  Wesley H. Huang Optimal line-sweep-based decompositions for coverage algorithms , 2001, Proceedings 2001 ICRA. IEEE International Conference on Robotics and Automation (Cat. No.01CH37164).

[55]  Esther M. Arkin,et al.  Angewandte Mathematik Und Informatik Universit at Zu K Oln Approximation Algorithms for Lawn Mowing and Milling Ss Andor P.fekete Center for Parallel Computing Universitt at Zu Kk Oln D{50923 Kk Oln Germany Approximation Algorithms for Lawn Mowing and Milling , 2022 .

[56]  Nancy M. Amato,et al.  MAPRM: a probabilistic roadmap planner with sampling on the medial axis of the free space , 1999, Proceedings 1999 IEEE International Conference on Robotics and Automation (Cat. No.99CH36288C).

[57]  Xueyin Lin,et al.  Fast road classification and orientation estimation using omni-view images and neural networks , 1998, IEEE Trans. Image Process..

[58]  Sylvia C. Wong,et al.  Natural Landmark Recognition using Neural Networks for Autonomous Vacuuming Robots , 2000 .

[59]  Howie Choset,et al.  Complete sensor-based coverage of unknown spaces: incremental construction of cellular decompositions , 2002 .

[60]  Bruce A. MacDonald,et al.  Natural landmark based localisation system using panoramic images , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[61]  Martin Held,et al.  On the Computational Geometry of Pocket Machining , 1991, Lecture Notes in Computer Science.

[62]  Howie Choset,et al.  Critical point sensing in unknown environments , 2000, Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065).

[63]  Howie Choset,et al.  Morse Decompositions for Coverage Tasks , 2002, Int. J. Robotics Res..

[64]  Ray A. Jarvis,et al.  DISTANCE TRANSFORM BASED PATH PLANNING FOR ROBOT NAVIGATION , 1994 .

[65]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[66]  U. Nehmzow QUANTITATIVE ANALYSIS OF ROBOT-ENVIRONMENT INTERACTION — ON THE DIFFERENCE BETWEEN SIMULATIONS AND THE REAL THING∗ , 2001 .

[67]  Luis Moreno,et al.  Path quality measures for sensor-based motion planning , 2003, Robotics Auton. Syst..

[68]  Robert E. Tarjan,et al.  Depth-First Search and Linear Graph Algorithms , 1972, SIAM J. Comput..

[69]  Mahmoud Tarokh,et al.  Manipulator path planning by decomposition: algorithm and analysis , 2001, IEEE Trans. Robotics Autom..

[70]  Michael Mayo,et al.  A Hybrid Approach to Finding Cycles in Hybrid Maps , 2003 .

[71]  Maja J. Mataric,et al.  Integration of representation into goal-driven behavior-based robots , 1992, IEEE Trans. Robotics Autom..

[72]  Sylvia C. Wong,et al.  Performance Metrics for Robot Coverage Tasks , 2002 .

[73]  Howie Choset,et al.  Coverage of Known Spaces: The Boustrophedon Cellular Decomposition , 2000, Auton. Robots.

[74]  Anthony Stentz,et al.  A comparative study between centralized, market-based, and behavioral multirobot coordination approaches , 2003, Proceedings 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003) (Cat. No.03CH37453).

[75]  Bernhard Nebel,et al.  Dynamic decentralized area partitioning for cooperating cleaning robots , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[76]  F. Clarke Generalized gradients and applications , 1975 .

[77]  Sylvia C. Wong,et al.  Creating Composite Images for Estimating the Effectiveness of Mobile Robot Coverage Algorithms , 2003 .

[78]  George Wolberg,et al.  Digital image warping , 1990 .

[79]  Lino Marques,et al.  Mobile pneumatic robot for demining , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[80]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[81]  Hong Zhang,et al.  Blind bulldozing: multiple robot nest construction , 2003, Proceedings 2003 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003) (Cat. No.03CH37453).

[82]  Anthony A. Maciejewski,et al.  Utilizing the topology of configuration space in real-time multiple manipulator path planning , 1994, Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS'94).

[83]  Howie Choset,et al.  Sensor-based Coverage of Unknown Environments: Incremental Construction of Morse Decompositions , 2002, Int. J. Robotics Res..

[84]  Rodney A. Brooks,et al.  A robot that walks; emergent behaviors from a carefully evolved network , 1989, Proceedings, 1989 International Conference on Robotics and Automation.

[85]  Francesco Mondada,et al.  Autonomous vacuum cleaner , 1997, Robotics Auton. Syst..

[86]  Thierry Siméon,et al.  Path coordination for multiple mobile robots: a resolution-complete algorithm , 2002, IEEE Trans. Robotics Autom..

[87]  Ralph L. Hollis,et al.  Distributed coverage of rectilinear environments , 2000 .

[88]  Sebastian Thrun,et al.  Learning Metric-Topological Maps for Indoor Mobile Robot Navigation , 1998, Artif. Intell..

[89]  Ray Jarvis,et al.  Covert path planning for autonomous robot navigation in known environments , 2003 .

[90]  John F. Canny,et al.  Constructing Roadmaps of Semi-Algebraic Sets I: Completeness , 1988, Artificial Intelligence.

[91]  Eric Haines,et al.  Point in Polygon Strategies , 1994, Graphics Gems.

[92]  Howie Choset,et al.  Exact cellular decompositions in terms of critical points of Morse functions , 2000, Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065).

[93]  Jayanthi Sivaswamy,et al.  Framework for practical hexagonal-image processing , 2002, J. Electronic Imaging.

[94]  Ming C. Lin,et al.  An opportunistic global path planner , 2005, Algorithmica.

[95]  Roland Siegwart,et al.  Hybrid simultaneous localization and map building: a natural integration of topological and metric , 2003, Robotics Auton. Syst..

[96]  Kurt Mehlhorn,et al.  LEDA: a platform for combinatorial and geometric computing , 1997, CACM.

[97]  Ralph L. Hollis,et al.  Contact sensor-based coverage of rectilinear environments , 1999, Proceedings of the 1999 IEEE International Symposium on Intelligent Control Intelligent Systems and Semiotics (Cat. No.99CH37014).